Re: [EAI] Thinking about requirements / downgrade
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [EAI] Thinking about requirements / downgrade



On Tue, 2009-09-15, Shawn Steele wrote:

> It would be much better to have a shared algorithmic method that we
> could all use (even if it was discouraged) rather than having different
> vendors create versions that are perhaps incompatible.  Can we create a
> draft (I can help if necessary) of a common mechanism or are there other
> objections besides the xn-- might conflict one?

I'm trying to remember all the reasons that were given when I advocated the
algorithmic approach several years ago.

The first was "Thou shalt not modify the left hand side!" Maybe all of the
issues brought up recently have softened this. Maybe not.

The second was "Some people expect to modify the left hand side!" For
example, the convention that a mailing list manager may be contacted by
appending "-request", like: <ima-request at ietf.org>. Punycode would mess
this up by appending stuff after the modified portion.

I don't remember if I bothered previously to suggest the ACE prefix: %%

The double percent would be using a character that historically had been used
for manual routing. Thus <McQuilWP%pobox.com at example.org> would be
initially sent to example.org which would forward it to me at
<McQuilWP at pobox.com>. This usage is now long deprecated, but may indicate
that the percent sign was considered reserved in left hand sides. Also a
double percent, like <%%punycode at example.com>, is perhaps very unlikely to
collide with any existing left hand side.

If the algorthmic method were guaranteed reversible, *any* ima-aware server
or client could downgrade or upgrade as it saw fit (e.g., to parse
embedded information in the left hand side) making sure only to downgrade
before passing it along to an ima-UNaware server.

-- 
Bill McQuillan <McQuilWP at pobox.com>


Note: Messages sent to this list are the opinions of the senders and do not imply endorsement by the IETF.